bacterial lipopolysaccharide regulates nociceptin expression in sensory neurons

Bacterial Lipopolysaccharide RegulatesNociceptin Expression in Sensory Neurons

Cristian Acosta and Alun Davies*

School of Biosciences, Cardiff, United Kingdom

Nociceptin/orphanin FQ (N/OFQ) is an opioid-relatedpeptide that is markedly up-regulated in sensory neu-rons in vivo following peripheral inflammation and playsa key role in pain physiology. To identify substancesthat up-regulate N/OFQ expression in sensory neurons,we carried out an in vitro screen using purified adultmouse dorsal root ganglion (DRG) neurons and identi-fied the potent proinflammatory agent bacterial lipo-polysaccharide (LPS) as a very effective inducer of N/OFQ. The robust response of these neurons to LPSenabled us to identify the components of a putativeneuronal LPS receptor complex. In contrast to theimmune system, where the functional LPS receptorcomplex is composed of CD-14 together with eitherMD-2 and TLR4 on myeloid cells or the homologousreceptors MD-1 and RP105 on mature B cells, DRGneurons express the unusual combination of CD-14,TLR4, and MD-1. Blocking antibodies against TLR4and MD-1 prevented induction of N/OFQ by LPS, and,in immunoprecipitation experiments, MD-1 coprecipi-tated with TLR4. Our findings suggest that LPS regu-lates N/OFN expression in sensory neurons via a novelcombination of LPS receptor components and demon-strate for the first time a direct action of a key initiatorof innate immune responses on neurons. VVC 2007 Wiley-Liss, Inc.

Key words: nociceptin/orphanin FQ; sensory neuron;lipopolysaccharide; receptor

Nociceptin/orphanin FQ (N/OFQ) is a 17-amino-acid opioid-related peptide that is produced by proteo-lytic cleavage of a precursor protein, prepronociceptin/orphanin FQ (PPN/OFQ; Okuda-Ashitaka and Ito,2000; Mogil and Pasternak, 2001). It is widely expressedin the central and peripheral nervous systems, includingin a subpopulation of small to medium-sized dorsal rootganglion (DRG) neurons (Pettersson et al., 2002; Mikaet al., 2003), and can act as a proalgesic (Inoue et al.,1998) or a proanalgesic (Inoue et al., 1999; Kolesnikovand Pasternak, 1999) peptide when administered periph-erally. Although expression of PPN/OFQ mRNA israpidly and markedly induced in DRG following carra-geenan-elicited paw inflammation (Andoh et al., 1997;Itoh et al., 2001), little is known about the factors thatcontrol its expression (Mogil and Pasternak, 2001). Toidentify regulators of N/OFQ expression, we screened

candidate molecules on the basis of their known proin-flammatory or pronociceptive functions and found toour surprise that bacterial lipopolysaccharide (LPS) is aparticularly effective inducer of N/OFQ in adult DRGneurons. This led to our discovery and characterizationof a neuronal LPS receptor complex.

LPS is a potent proinflammatory agent that pro-motes hyperalgesia and pain (Watkins et al., 1994;Kanaan et al., 1996; Reeve et al., 2000). In the immunesystem, the core components of the most thoroughlycharacterized LPS receptor complex are CD14, Toll-likereceptor 4 (TLR4), and myeloid differentiation protein-2 (MD-2). CD14 is a glycoprotein occurring in solubleform in the circulation and as a GPI membrane-anch-ored form on myelomonocytic cells (Wright et al.,1990) that concentrates LPS and presents it to a cell sur-face TLR4–MD-2 complex (Haziot et al., 1998; Muroiet al., 2002) TLR4 is a transmembrane protein whosecytoplasmic domain contains a signal transducing Toll/IL-1 receptor (TIR) homology domain (Medzhitovet al., 1997) MD-2 is a secreted glycoprotein that physi-cally associates with the extracellular domain of TLR4(Shimazu et al., 1999; Viriyakosol et al., 2001) and isessential for LPS signaling via TLR4 in macrophages,dendritic cells, and some B cells (Schromm et al., 2001;Visintin et al., 2001; Nagai et al., 2002a). In addition tothe TLR4–MD-2 complex, macrophages, dendritic cells,and mature B cells express a related LPS receptor com-plex comprising RP105, which is structurally similar toTLR4 but lacks a cytoplasmic TIR domain, and MD-1,which is structurally similar to MD-2 (Fugier-Vivieret al., 1997; Miura et al., 1998; Miyake et al., 2000;Zarember and Godowski, 2002). Although the responsesof macrophages and dendritic cells from RP1052/2

or MD-12/2 mice are not impaired to LPS, absence ofeither RP105 or MD-1 leads to reduced responses of Bcells to LPS (Chan et al., 1998; Nagai et al., 2002b).

Contract grant sponsor: Wellcome Trust.

*Correspondence to: Alun Davies, School of Biosciences, Biomedical

Building, Museum Avenue, Cardiff CF10 3US, United Kingdom.

E-mail: [email protected]

Received 19 April 2007; Revised 29 August 2007; Accepted 1

September 2007

Published online 20 November 2007 in Wiley InterScience (www.

interscience.wiley.com). DOI: 10.1002/jnr.21565

Journal of Neuroscience Research 86:1077–1086 (2008)

' 2007 Wiley-Liss, Inc.

Thus, the available evidence suggests the existence oftwo different LPS receptor complexes in cells of theimmune system that comprise CD-14 together with ei-ther MD-2 plus TLR4 or MD-1 plus RP105. Our stud-ies suggest that a distinctive combination of LPS receptorcomponents is necessary for the effects of LPS onneurons.

MATERIALS AND METHODS

Cell Culture and Animals

Sensory neurons from DRG of embryonic, newborn,and adult CD-1 mice were isolated as described previously(Acosta and Lopez, 1999; Acosta et al., 2001). Briefly, DRGwere enzymatically dissociated by incubation for 15–45 min(depending on age) with 0.25% trypsin and 1.25% collagenaseat 378C (Worthington, Lakewood, NJ). After washing withEagle’s minimal essential medium containing 10% fetal bovineserum (FBS), the ganglia were gently triturated, and theresulting cell suspension was pelleted by centrifugation at2,000 rpm for 5 min. To remove most of the nonneuronalcells, the pellet was resuspended in HBSS, layered on a 20%Percoll gradient, and centrifuged at 2,500 rpm for 6–8 min.The supernatant containing the nonneuronal cells was dis-carded, and the pellet of enriched neurons was suspended inF14-based defined medium and plated in poly-dl-ornithine/laminin-coated 35-mm tissue culture dishes (Davies et al.,1993). LPS was added 1 hr after plating, and 5–10 lM b-ara-binofuranosylcytosine (b-ARAC) was added 6 hr after platingto eliminate dividing fibroblasts. The Percoll gradient and b-ARAC treatment resulted in cultures that were more than95% neurons after 1 day in vitro (DIV). For immunocyto-chemistry, the cells were plated on poly-dl-ornithine/laminin-coated 12-mm-diameter glass coverslips (Bellco Glass,Vineland, NJ). HEK293 cells (ECACC No. 85120602) wereobtained from the European Collection of Cell Cultures(Wiltshire, United Kingdom). These were maintained andpropagated as described elsewhere (Kolanus et al., 1996).

Immunocytochemistry

The cultures were fixed with 4% paraformaldehyde plus4% sucrose for 20 min at 378C and washed with phosphate-buffered saline (PBS). For detection of intracellular proteins(bIII-tubulin and N/OFQ), the cells were permeablized with0.2% Triton X-100 in PBS for 5 min at room temperature,and nonspecific binding sites were blocked with 5% bovineserum albumin (BSA) for 2 hr. Permeablization was omittedfor detection of cell surface antigens (TLR4, MD-1, andRP105). The cultures were incubated with primary antibodiesovernight at 48C in 1% BSA at the dilutions indicated inResults. The cultures were then washed three times withPBS, and the cells were incubated for 1 hr at room tempera-ture with the corresponding secondary antibody, which waslabelled with either fluorescein isothiocyanate (FITC) orrhodamine. The preparations were mounted in FluorSave(Calbiochem, La Jolla, CA), and images were digitallyacquired and the level of cellular fluorescence quantified usingZeiss Axiovert and LSM510 confocal microscopes. A series ofpreliminary experiments established the appropriate dilutions

of primary antibodies that permitted labelled neurons to beclearly recognized without any significant background stain-ing. To identify which neurons expressed N/OFQ, a compar-ison between two ratios was undertaken. The first ratio wasthe intensity of the C-17 Ab fluorescent signal to that of thebIII-tubulin signal. The second ratio was the background sig-nal (obtained from cultures in which cells were pretreatedwith the blocking peptide for C-17 Ab at 378C for 6 hr) tothat of the bIII-tubulin signal. Where the former ratioexceeded the later by more than 50%, a particular neuron wasconsidered as positively labelled for N/OFQ.

In Vitro LPS Binding

Purified P60 DRG neurons were plated in 200-ll videomicroscopy chambers (Acosta et al., 2001) and incubated withnerve growth factor (NGF). After 24 hr, the cells werewashed and exposed for 30 min at 378C to 1 lg/ml LPS sero-type 055:B5 tagged with Alexa-568 (Molecular Probes,Eugene, OR) alone or in combination with 50ng/mlrecombinant lipid-binding protein (Biometec, Greifswald,Germany). After being washed with PBS, the neurons werestudied via confocal microscopy.

Western Blotting

The expression of TLR4, MD-1, RP105, and MD-2 intissues and cultured DRG neurons of P60 mice was examinedby Western blot using standard protocols (Acosta et al., 2001).Tissues and harvested neurons (three to five dishes per condi-tion) were homogenized at 48C in RIPA 1 3 buffer supple-mented with Complete Protease-Inhibitor Cocktail (RocheMolecular Biochemicals, South Essex, United Kingdom).Samples were centrifuged at 14,000 rpm for 15 min at 48C;the supernatant was recovered, centrifuged again, and kept at2208C. The proteins were quantified using colorimetricmethods, and 20-lg protein samples were run in each lane ofSDS-PAGE gels (normally 7.5% gels, and 12% gels for detect-ing MD-1 or MD-2). Proteins were blotted onto PVDFmembranes (Immobilon-P), and proteins were detected usingthe ECL Plus staining system (Amersham Biosciences, Amer-sham, United Kingdom).

Semiquantitative RT-PCR

The relative levels of CD-14, TLR4, MD-1, MD-2,and RP105 mRNAs in cultured DRG neurons were deter-mined by RT-PCR. Total mRNA was obtained using theQiagen RNeasy Mini kit (Qiagen, East Sussex, United King-dom). cDNA was synthesized from 5 lg total RNA with ran-dom hexamers and SuperScript H2 (Invitrogen, Gaithersburg,MD) in the presence of a RNAse inhibitor (RNAguard;Amersham Biosciences) for 50 min at 378C. Five microliterscDNA was used as a template for PCR amplification in a 25-ll reaction volume containing 1 3 PCR buffer, 100 nMdNTPs,100 nM of each primer and taq DNA polymerase(Helena BioSciences Europe, Sunderland, United Kingdom).The reactions were as follows: CD-14, 32 cycles of 948C for1 min, 568C for 1 min, and 688C for 1 min; TLR4, 30 cyclesof 948C for 1 min, 548C for 1 min, and 688C for 1 min;MD-1, 30 cycles of 948C for 1 min, 538C for 75 sec, and

1078 Acosta and Davies

Journal of Neuroscience Research

688C for 1 min; MD-2, 32 cycles of 948C for 1 min, 52.58Cfor 45 sec, and 688C for 75 sec; RP105, 33 cycles of 948C for1 min, 538C for 75 sec, and 688C for 75 sec. All had a final10-min 688C extension. In all experiments, amplification ofglyceraldehyde-3-phosphate dehydrogenase (GAPDH) was runin parallel to normalize samples. Primers were obtained fromMWG Biotech (London, United Kingdom). The sequencesand sizes of the products (in parenthesis) were: CD-14, senseGCCCTCTCCACCTTAGACCT, antisense TCAGCCCAGTGAAAGACAGA (277); TLR4, sense GAGCCGTTGGTGATCTTTG, antisense TGCCGTTTCTTGTTCTTCC(274); MD-1, sense CCTATCCCCTTTGTGAGGAG, anti-sense CTTGGTTATCAGTGGTTCTTGC (282); MD-2,sense GACGCTGCTTTCTCCCATA, antisense CTTACGCTTCGGCAACTCTA (250); RP105, sense TTTCCCCTCCCCTTACTCACC, antisense CTTTGAATGCCTCCGTCTTG (252). Gels were stained with ethidium bromide.

Expression Vectors

Expression vectors for murine MD-1 (pUNO-mMD1)and TLR4 (pUNO-mTLR4) were obtained from Invitrogen(San Diego, CA). To make fluorescent-tagged versions ofboth proteins, full-length cDNAs for MD-1 and TLR4 werecloned into the C-terminus of pDsRed2-C1 (BD BiosciencesClontech, Oxford, United Kingdom), generating the plasmidspTLR4-DsREd2-C1 and pMD1-DsRed2-C1, which werechecked by sequencing (MWG Biotech).

Transfection and Immunoprecipitation

HEK293 cells were grown at 70–80% confluence in60 mm Petri dishes in MEM supplemented with 2% fetal bo-vine srum (FBS). Prior to transfection, the serum containingmedium was replaced with 5 ml fresh N2 medium. A mixtureof 200 ll LyoVec reagent (Invitrogen, San Diego, CA) plusplasmid DNA preincubated at RT for 30 min was added tothe culture medium. Highly purified, endotoxin-free plasmidDNA was obtained by using the Qiagen endo-free kit. Thefollowing quantities of plasmids were used per transfection:4 lg pUNO-mMD1 or pUNO-mTLR4, 2 lg pDsRed2-C1(vector), 4 lg pMD1-DsRed2-C1, or pTLR4-DsRed2-C1and 4 lg each pUNO-mMD1 plus pTLR4-DsRed2-C1. ForWestern blot or immunoprecipitation (IP), cells were har-vested 24 or 72 hr after transfection. For IP, the lysis buffercomprised PBS containing 5 mM EDTA, 0.5% NP-40, 0.5%DOC-Na, and Complete Protease Inhibitor (Roche), pH 7.3.After incubation with 1 ml lysis buffer for 15 min at 48C, thecells were scraped off the dish, vortexed for 30 sec, passedthrough a fine-bore needle three times, and sonicated for2 min. The samples were centrifuged at 48C for 15 min at13,000 rpm, and the supernatant was kept on ice. To preclearthe samples, 20 ll of protein-G-agarose beads (AmershamBiosciences, United Kingdom) was added to each sample androtated for 1 hr at 48C. After 1 min of centrifugation, thesupernatant was recovered, and the beads were kept to checkfor nonspecific binding. Two and one-half microliters of a1:10 dilution of rabbit anti-DsRed antibody (Clontech) wasadded to the supernatant and incubated under continuousrotation overnight at 48C. On the next day, 50 ll protein-G-

agarose beads was added to each tube and incubated at 48Cfor 2 hr. The beads were then washed five times with 1 mlPBS and resuspended in 30 ll Laemmli buffer, boiled at1008C for 3 min, and run on a 4–15% premade Bio-Rad gelunder denaturing conditions. Western blotting was carried outas described above.

Antibodies and Reagents

The goat polyclonal anti-mouse N/OFQ (C-17) and ratmonoclonal IgG2a anti-mouse RP105 (RP14) were fromSanta Cruz Biotechnologies (Autogen Bioclear, Wiltshire,United Kingdom). The anti-mouse MD-1 monoclonal, cloneMD113 was obtained from Cell Sciences, Inc. (Norwood,MA). Affinity-purified rabbit anti-mouse MD-1 polyclonalantibody (raised against a KLH-conjugated synthetic peptidecorresponding to amino acids 112–125 of human MD-1) wasfrom Imgenex (San Diego, CA). The immunoprecipitating/blocking rat monoclonal anti-TLR4, clone MTS510, and therabbit anti-mouse TLR4 polyclonal antibody were both fromAbcam Ltd. (Cambridge, United Kingdom). All secondaryantibodies were acquired from Jackson Immunochemicals(Stratech Scientific Ltd., Cambridgeshire, United Kingdom).All antibodies used in cell culture experiments were first dia-lyzed against culture medium to eliminate the sodium azideused in storage. LPS (Escherichia coli serotype 055:B5), themonoclonal mouse anti-b-tubulin isotype III (cloneSDL.3D10), antithrombin III (AT III) and all other reagentswere obtained from Sigma (Dorset, United Kingdom). NGFwas obtained from Merck Biosciences (Nottingham, UnitedKingdom).

RESULTS

Expression and Regulation of N/OFQ inCultured DRG Neurons

To investigate the factors that act directly on sen-sory neurons to regulate N/OFQ expression, we estab-lished highly enriched dissociated neuron cultures(>95% neurons, confirmed by bIII-tubulin staining)from adult mouse (postnatal day 60; P60) DRG. Theability of adult DRG neurons to survive in culture with-out added neurotrophic factors (Lindsay, 1988) enabledus to compare N/OFQ expression in nonstimulatedcontrol cultures and in cultures treated with candidateinducers. We restricted our analysis to the expression ofN/OFQ protein rather than the expression of mRNAencoding the N/OFQ sequence, because N/OFQ isproduced by proteolytic cleavage of a precursor protein,PPN/OFQ, from which other peptides such as nocista-tin are also generated (Okuda-Ashitaka and Ito, 2000).We used immunocytochemistry to quantify the propor-tion of DRG neurons expressing N/OFQ in unstimu-lated control cultures and cultures treated with candidateneurotrophic factors and other molecules chosen becauseof their known proinflammatory or pronociceptivefunctions.

Our initial screen of candidate molecules revealedthat bacterial LPS was a particularly effective inducer ofN/OFQ expression. Figure 1A shows that approximately

LPS and Nociceptin Expression 1079


25% of P60 neurons were N/OFQ-positive in control,unstimulated cultures after 24 hr of incubation and thatthis proportion did not significantly change up to 96 hrin culture. LPS doubled the number of N/OFQ-positiveneurons in P60 cultures after 24 hr of treatment (Fig.1A). Dose-response analysis showed that the concentra-tion of LPS used in these experiments (0.5 lg/ml) wasmaximally effective for N/OFQ expression (data notshown). The number of N/OFQ-positive neuronsin LPS-stimulated cultures returned to basal levels after48 hr of treatment with LPS (Fig. 1A).

Because LPS has been reported to cause death ofCNS neurons in vivo and in vitro (Kim et al., 2000;Eklind et al., 2001; Nguyen et al., 2002; Arai et al.,2003), we investigated whether the magnitude of theLPS effect on N/OFQ expression is masked by a detri-mental effect on the viability of DRG neurons. Figure1B shows that 10 lg/ml LPS, a concentration that is 20-fold higher than that required to stimulate N/OFQexpression maximally, did not affect the survival ofP60 DRG neurons. Virtually all neurons survived inthe presence of this high concentration of LPS. The typ-ical appearance of P60 DRG neurons incubated withand without LPS and stained for N/OFQ is shown inFigure 2A.

To investigate the regulation of N/OFQ expressionin fetal DRG neurons, it was necessary to include NGF inthe medium to sustain the survival of these neurons. Theproportion of N/OFQ-positive neurons in cultures of E18DRG neurons after 24 hr of incubation with NGF alone(Fig. 1C) was similar to the proportion of N/OFQ-posi-tive neurons in P60 control cultures (Fig. 1A). As in adultcultures, the proportion of N/OFQ-positive neurons inE18 cultures was increased over twofold by the additionto LPS to the medium (Fig. 1C). Figure 1B shows thatvirtually all NGF-supported E18 neurons survived with 10lg/ml LPS, demonstrating that the magnitude of the effectof LPS on N/OFQ expression was unaffected by differen-ces in neuronal survival. In the absence of NGF, almost80% of the neurons died within 24 hr (data not shown).The typical appearance of E18 DRG neurons incubatedwith and without LPS and stained for N/OFQ is shownin Figure 2B.

DRG Neurons Bind LPS In Vitro

We have shown that purified DRG neurons fromfetal and adult mice respond to LPS by up-regulating theexpression of N/OFQ. These findings imply that LPSacts on DRG neurons by a receptor-mediated mecha-nism. The first step in substantiating this assumption wasto show that LPS binds to DRG neurons. Although LPScan bind stably to soluble and membrane-associatedforms of the CD14 LPS receptor and elicit cellresponses, LPS binding to CD14 is catalytically acceler-ated by a trace plasma protein, lipopolysaccharide-bind-ing protein (LBP; Hailman et al., 1994; Yu and Wright,1996). For this reason, we incubated cultures of purifiedadult DRG neurons with fluorescently tagged LPS alone

Fig. 1. LPS enhances N/OFQ expression in embryonic andadult DRG neurons. Purified DRG neurons from P60 adultand E18 embyros were incubated in defined medium alone (con-trol) or medium containing NGF (50 ng/ml) or LPS (0.5 lg/ml,A, C; or 10 lg/ml, B) or NGF and LPS combined. A and C showthe mean percentage 6 SEM of bIII-tubulin-positive cells that wasimmunostained by C-17 N/OFQ antibody. B shows the percent-age survival 6 SEM. Each bar represents data obtained frombetween three and six independent experiments. **P < 0.05,statistical comparison with 24 hr control (A) or NGF alone (C), t-tests.



and with tagged LPS plus LBP. The neurons were incu-bated with these reagents for 30 min and were observedunder confocal microscopy immediately after washing.Figure 3 shows that, although tagged LPS bound to someneurons, the labelling intensity was low. However, the in-tensity of labelling greatly increased in the presence ofLBP. Approximately 37% of the neurons were clearlylabelled by fluorescently tagged LPS in the presence ofLBP. These observations demonstrate that DRG neuronsbind LPS in vitro. Although LBP increased the intensitywith which neurons were labelled with fluorescentlytagged LPS, the addition of LBP to the culture mediumdid not further enhance the number of N/OFQ-positiveneurons after 1 day in vitro (data not shown). Separateexperiments had shown that the LPS dose used through-

out the study was maximally effective for enhancing N/OFQ expression, so, despite the enhanced LPS bindingbrought about by LBP, there was no additional effect onN/OFQ expression.

LPS Receptor Components Expressed by AdultDRG Neurons

We used a combination of semiquantitative RT/PCR, Western blotting and immunocytochemistry to as-certain whether purified P60 DRG neurons express thecore components of the LPS receptor complexesdescribed in myeloid cells and B cells. Figure 4A showsthat purified P60 DRG neurons grown for 24 hr withand without LPS in the culture medium exhibited robustexpression of TLR4, MD-1, and CD14 mRNAs. Incontrast, RP105 mRNA was undetectable, and negligi-ble MD-2 amplification products were obtained athigher cycle number from these neurons, yet robustproducts for both RP105 and MD-2 mRNAs wereamplified from a similar amount of total RNA extractedfrom spleen. RT/PCR was carried out for the house-keeping GAPDH mRNA in parallel, to facilitate com-parison of the relative levels of receptor transcripts inpurified neurons and spleen. Western blotting revealedthat purified DRG neurons grown for 24 hr with andwithout LPS in the culture medium expressed significantlevels of TLR4 and MD-1 proteins but lacked RP105protein and had barely detectable amounts of MD-2protein, although all four proteins were clearly presentin the same amount of total protein extracted fromspleen (Fig. 4B). These findings suggest that adult DRGneurons express the high-affinity LPS receptor CD14,together with the signal-transducing LPS receptorTLR4, but express only very low levels of the TLR4-associated coreceptor MD-2. However, the neurons

Fig. 3. LPS binds to DRG neurons in culture. Representative confocalimages of live P60 DRG neurons incubated for 30 min at 378C with afluorescently tagged LPS (LPS-Alexa) 24 hr after plating. The left pho-tograph shows neurons exposed to 1 lg/ml LPS-Alexa alone, whereasthe right photograph shows neurons exposed to 1 lg/ml LPS-Alexaplus 50 ng/ml LBP. Yellow arrows indicate typical examples of neuronswith bound LPS, whereas white arrows indicate typical examples ofneurons without significant LPS binding. Scale bar 5 50 lm. [Colorfigure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Fig. 2. N/OFQ staining of DRG neurons incubated with and with-out LPS. Matching images of p60 (A) and E18 (B) DRG neuronsdoubly labelled for bIII-tubulin (left panels) and N/OFQ (right pan-els) after 24 hr of incubation with and without 0.5 lg/ml LPS. E18cultures additionally received 50 ng/ml NGF to sustain neuronal sur-vival. Scale bars 5 100 lm. [Color figure can be viewed in theonline issue, which is available at www.interscience.wiley.com.]



clearly express the homologous glycoprotein MD-1 butnot the TLR4-related transmembrane protein RP105with which MD-1 has been shown to associate inmature B-cell lymphocytes.

These findings raised the possibility that TLR4-MD-1 may form part of a novel receptor complex forLPS on neurons. To explore this possibility further, weused immunocytochemistry to ascertain the proportion

of DRG neurons expressing TLR4 and MD-1 andwhether these receptors are coexpressed on the sameneurons. The representative field of P60 DRG neuronsimmunostained for TLR4 and MD-1 illustrated in Fig-ure 5A shows that many neurons clearly coexpress bothreceptors. Quantification of the proportion of neuronslabelled by anti-TLR4 and anti-MD-1 revealed that justover 40% of the neurons expressed TLR4 and approxi-

Fig. 4. DRG neurons express LPS receptor components. A: Repre-sentative gels of RT/PCR amplification products for TLR4, MD-1,MD-2, RP105, and CD-14 mRNAs relative to GAPDH mRNA inpurified P60 DRG cultures grown for 24 hr in defined medium inthe absence of LPS (Con) or with 0.5 lg/ml LPS. B: Representative

Western blot analysis of TLR4, MD-1, MD-2, and RP105 proteinin purified P60 DRG neurons cultured for 24 hr with 0.5 lg/mlLPS. Approximately 20 lg of total protein were run in each lane. Inall experiments, spleen tissue (Sp) was used as a positive control.



mately 70% expressed MD-1 (Fig. 5B). Importantly, therewas no significant difference between the number ofTLR4-labelled and TLR4/MD-1 double-labelled neurons(45.8% vs. 41.7%), suggesting that all TLR4-positive neu-rons coexpress MD-1. These observations raised the possi-bility that TLR4/MD-1 may form or be components of afunctional receptor for LPS on neurons.

MD-1 Precipitates With TLR4

Given the physical association of MD-2 withTLR4 (Shimazu et al., 1999; Viriyakosol et al., 2001),we used immunoprecipitation to investigate whetherinteraction between MD-1 and TLR4 is possible.Because harvesting enough protein from purified cul-tures of DRG neurons proved extremely difficult, transi-ently transfected HEK293 cells that do not normallyexpress TLR4, MD-1, or MD-2 (Chow et al., 1999;Latz et al., 2002; Kurt-Jones et al., 2004) were employedin these experiments. Western analysis revealed that cellstransfected with pUNO-TLR4 expressed a single band

Fig. 6. Association of MD-1 with TLR4 and requirement of boththese proteins for a functional response of DRG neurons to LPS.A, B: Immunoblots of total protein extracted from HEK293 cells24 and 72 hr after transfection with pUNO-TLR4 (A) andpUNO-MD-1 (B) and stained with anti-TLR4 and anti-MD-1antibodies, respectively. For comparison with the level of endoge-nous expression of TLR4 and MD-1, total protein from DRG andspleen (Sp) was also run on these gels. C: Proteins from HEK293cells 72 hr after transfection with an empty pDsRed2-C1 vector(pVc), pTLR4-DsRed-2-C1 (T4), and pTLR4-DsRed-2-C1 pluspUNO-MD-1 (T4 1 M) were immunoprecipitated with an anti-DsRed antibody and blotted with an anti-MD-1 antibody.Approximately 20 lg of total protein was run in all lanes for blotsA, B, and C. D: Bar chart of mean percentage 6 SEM of bIII-tubulin-positive cells immunostained by C-17 N/OFQ antibody inP60 DRG cultures that were pretreated for 5 min with 2 lg/ml ofAT-III or preincubated for 1 hr with 1:100 dilutions of anti-TLR4(MTS510) or anti-MD-1 (MD-113) antibodies, followed by 24 hrof incubation with or without 0.5 lg/ml LPS. Each bar representsdata obtained from three independent experiments. **Statisticallysignificant differences between the treatments and the control (P <0.05, ANOVA).

Fig. 5. DRG neurons are labelled with anti-MD-1 and anti-TLR4antibodies. A: Representative confocal images of cultured P60 DRGneurons doubly immunostained for MD-1 and TLR4. Arrows indi-cate typical examples of robust coexpression. B: Bar chart of the per-centage of P60 DRG neurons clearly immunopositive for MD-1,TLR4, and MD-1 plus TLR4 after 24 hr of culture in defined me-dium (mean and SEM of data from three independent experiments).Scale bar 5 200 lm. [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.]



of �105 kDa corresponding to TLR4, and cells trans-fected with pUNO-MD-1 expressed bands of �22 kDaand �25 kDa corresponding to MD-1, as previouslydescribed (Miura et al., 1998) In both cases, peak proteinexpression was achieved by 72 hr after transfection (Fig.6A,B). Because the commercially available TLR4 immu-noprecipitating antibody (MTS510) immunoprecipitatesTLR4 only if it is associated with MD-2 (Akashi et al.,2000b), we constructed a TLR4 fusion protein carryingthe DsRed fluorescent tag by subcloning full-lengthmouse TLR4 cDNA into the pDsRed2-C1 Clontechvector. As with the native TLR4 protein, Western blot-ting revealed that expression of the TLR4-DsRed fusionprotein in HEK293 cells peaked 72 hr after transfection,and transfected cells exhibited clear cell surface fluores-cence (not shown). To demonstrate association of TLR4and MD-1, an anti-DsRed antibody was used to immu-noprecipitate the TLR4-DsRed fusion protein fromHEK293 cells that had been transfected with pTLR4-DsRed-2-C1 with and without pUNO-MD-1. Sampleswere run under denaturing conditions and blotted withan anti-MD-1 antibody. A clear band of �22 kDa and afainter one of �25 kDa were observed from cells trans-fected with pTLR4-DsRed2-C1 plus pUNO-MD1.Cells transfected with pTLR4-DsRed2-C1 alone orwith the empty pDsRed2-C1 vector exhibited onlyfaint, nonspecific bands (Fig. 6C). Similar, faint, nonspe-cific bands were observed in protein extracted fromnontransfected HEK293 cells blotted with the anti-MD-1 antibody (not shown). These results suggest thatTLR4 and MD1 are physically associated when coex-pressed in HEK293 cells.

Blockade of TLR4 and MD-1 PreventsLPS Effect on N/OFQ Expression

To explore the functional relevance of TLR4 andMD-1 in mediating the response of DRG neurons toLPS, we investigated the effects of inhibitors of TLR4-and MD-1-dependent LPS signalling. Transient exposureof TLR4-expressing cell lines to the serine protease in-hibitor antithrombin III (AT-III) selectively blocks acti-vation of nuclear factor-jB by LPS (Mansell et al.,2001). Exposing P60 DRG neurons to AT-III for 5 minprior to LPS treatment completely prevented the induc-tion of N/OFQ by LPS (Fig. 6D). Likewise, pretreatingP60 DRG neurons for 1 hr with a rat monoclonal anti-body (MTS510) that inhibits TLR4-mediated release ofcytokines (Akashi et al., 2000a) inhibited N/OFQinduction by LPS (Fig. 6D). Similar pretreatment of P60DRG neurons with the MD-113 function-blockinganti-MD-1 mouse monoclonal antibody (Miyake et al.,1998; Yazawa et al., 2003) also inhibited induction ofN/OFQ expression in these neurons by LPS (Fig. 6D).

DISCUSSION

We have demonstrated for the first time that neu-rons possess receptors for a key initiator of innateimmune responses. We have shown that LPS binds

directly to adult DRG neurons and that these neuronsexpress components of LPS receptor complexes identi-fied in the immune system. These receptors are func-tional, because specific blocking antibodies to a distinc-tive combination of receptor components inhibit theresponse of the neurons to LPS. These findings providenew insights into the cooperative response of theimmune and nervous systems to infection and haveimplications for our understanding of the complexities ofnociceptive responses to inflammation. Our study hasalso identified the first factor that is capable of directlyregulating the level of the pain modulating peptide N/OFQ in adult DRG neurons. In purified, dissociatedcultures, half of the neurons expressed N/OFQ after24 hr of treatment with LPS, compared with only aquarter of the neurons in unstimulated cultures. Thisincrease in the proportion of N/OFQ-positive neuronsmight have been due to a change in either the expres-sion or the processing of the N/OFQ precursor proteinPPN/OFQ. Although previous work has shown thatLPS induces calcitonin gene-related peptide (CGRP)release in dissociated neonatal rat DRG cultures (Houand Wang, 2001), it was unclear whether LPS acteddirectly on the neurons or indirectly affected CGRPrelease by acting on nonneuronal cells, because non-neuronal cells were not removed from the cultures.There is no such ambiguity in our study, insofar as ourcultures were virtually devoid of nonneuronal cells.

Our findings suggest that the LPS receptor com-plex expressed on neurons differs from previously char-acterized LPS receptor complexes in the immune system.Purified adult DRG neurons expressed MD-1 andTLR4 mRNA and protein at levels comparable to thosein spleen (relative to GAPDH mRNA and total protein,respectively). However, the MD-1-associated coreceptorRP105 that forms an integral component of the LPS re-ceptor complex in mature B cells (Miyake et al., 2000)was not detectable at either the mRNA or protein levelsin these neurons. Also, the TLR4-associated coreceptorMD-2 that forms an integral component of the LPS re-ceptor complex in myeloid cells (Miyake et al., 2000)was barely detectable at the mRNA and protein levels inthese neurons. Our finding that virtually all TLR4-expressing neurons (about 40% of the neurons in adultDRG cultures) coexpressed MD-1 and our demonstra-tion that function-blocking antibodies against TLR4 andMD-1 selectively prevented enhanced expression of N/OFQ by LPS suggest that TLR4 and MD-1 are bothrequired for the response of DRG neurons to LPS. Therequirement for TLR4 was additionally supported by theselective inhibition of N/OFQ expression by LPS inDRG neurons treated with the TLR4 antagonist AT-III. Furthermore, our demonstration that MD-1 wasprecipitated by a chimeric TLR4 protein in the mem-brane-associated fraction from HEK293 cells cotrans-fected with these proteins suggests that TLR4 and MD-1 have the ability to interact physically in the absence ofMD-2, providing additional support for the idea thatMD-1 and TLR4 may form part of a receptor complex



when present together in cells. Our finding that TLR4and MD-1 can interact when coexpressed contrasts witha previous report in which a secreted flag-taggedrecombinant form of MD-1 failed to interact in solutionwith a recombinant form of soluble TLR4 (Re andStrominger, 2002). However, in our experiments, westudied the interaction of the physiologically relevant,nonsoluble, membrane form of TLR4 and used a non-modified form of mouse MD-1. Although our resultsshow that TLR4 and MD-1 are expressed at relativelyhigh levels in DRG neurons and are both necessarycomponents of a functional LPS receptor complex inthese cells, we cannot exclude the possibility that MD-2is also a component of this receptor complex or isrequired for the assembly of a functional LPS receptorcomplex because, low levels of MD-2 were detected inDRG neurons. There is evidence, for example, thatMD-2 is required for the movement of TLR4 from theGolgi apparatus to the cell surface (Nagai et al., 2002a),and it is possible that MD-2 performs at least this func-tion in neurons.

In summary, we have shown that LPS directly reg-ulates the expression of the nociceptive peptide N/OFQand that a distinctive combination of LPS receptors isnecessary for LPS to exert its effects in neurons. Thesefindings imply that gram-negative bacterial infections arecapable of directly influencing sensory neurons in addi-tion to the indirect effects of bacterial-induced inflam-mation on the sensitization of these neurons. Our find-ings provide an intriguing example of a set of receptorsshared between the immune and the nervous systemsthat modulate or coordinate responses of both systems toa common stimulus.

REFERENCES

Acosta CG, Lopez HS. 1999. Delta opioid receptor modulation of several

voltage-dependent Ca21 currents in rat sensory neurons. J Neurosci

19:8337–8348.

Acosta CG, Fabrega AR, Masco DH, Lopez HS. 2001. A sensory neuron

subpopulation with unique sequential survival dependence on nerve

growth factor and basic fibroblast growth factor during development.

J Neurosci 21:8873–8885.

Akashi S, Ogata H, Kirikae F, Kirikae T, Kawasaki K, Nishijima M,

Shimazu R, Nagai Y, Fukudome K, Kimoto M, Miyake K. 2000a. Reg-

ulatory roles for CD14 and phosphatidylinositol in the signaling via toll-

like receptor 4-MD-2. Biochem Biophys Res Commun 268:172–177.

Akashi S, Shimazu R, Ogata H, Nagai Y, Takeda K, Kimoto M, Miyake

K. 2000b. Cutting edge: cell surface expression and lipopolysaccharide

signaling via the toll-like receptor 4-MD-2 complex on mouse perito-

neal macrophages. J Immunol 164:3471–3475.

Andoh T, Itoh M, Kuraishi Y. 1997. Nociceptin gene expression in rat

dorsal root ganglia induced by peripheral inflammation. Neuroreport

8:2793–2796.

Arai K, Wood JP, Osborne NN. 2003. Beta-adrenergic receptor agonists

and antagonists counteract LPS-induced neuronal death in retinal cul-

tures by different mechanisms. Brain Res 985:176–186.

Chan VW, Mecklenbrauker I, Su I, Texido G, Leitges M, Carsetti R,

Lowell CA, Rajewsky K, Miyake K, Tarakhovsky A. 1998. The mo-

lecular mechanism of B cell activation by toll-like receptor protein

RP105. J Exp Med 188:93–101.

Chow JC, Young DW, Golenbock DT, Christ WJ, Gusovsky F. 1999.

Toll-like receptor-4 mediates lipopolysaccharide-induced signal trans-

duction. J Biol Chem 274:10689–10692.

Davies AM, Lee KF, Jaenisch R. 1993. p75-Deficient trigeminal sensory

neurons have an altered response to NGF but not to other neurotro-

phins. Neuron 11:565–574.

Eklind S, Mallard C, Leverin AL, Gilland E, Blomgren K, Mattsby-

Baltzer I, Hagberg H. 2001. Bacterial endotoxin sensitizes the immature

brain to hypoxic-ischaemic injury. Eur J Neurosci 13:1101–1106.

Fugier-Vivier I, de Bouteiller O, Guret C, Fossiez F, Banchereau J, Mat-

tei MG, Ait-Yahia S, Garcia E, Lebecque S, Liu YJ. 1997. Molecular

cloning of human RP105. Eur J Immunol 27:1824–1827.

Hailman E, Lichenstein HS, Wurfel MM, Miller DS, Johnson DA, Kel-

ley M, Busse LA, Zukowski MM, Wright SD. 1994. Lipopolysaccha-

ride (LPS)-binding protein accelerates the binding of LPS to CD14. J

Exp Med 179:269–277.

Haziot A, Hijiya N, Goyert SM. 1998. Role of CD14 in infection: stud-

ies in CD14-deficient mice. Prog Clin Biol Res 397:255–260.

Hou L, Wang X. 2001. PKC and PKA, but not PKG mediate LPS-

induced CGRP release and [Ca21]i elevation in DRG neurons of neo-

natal rats. J Neurosci Res 66:592–600.

Inoue M, Kobayashi M, Kozaki S, Zimmer A, Ueda H. 1998. Nocicep-

tin/orphanin FQ-induced nociceptive responses through substance P

release from peripheral nerve endings in mice. Proc Natl Acad Sci U S A

95:10949–10953.

Inoue M, Shimohira I, Yoshida A, Zimmer A, Takeshima H, Sakurada

T, Ueda H. 1999. Dose-related opposite modulation by nociceptin/

orphanin FQ of substance P nociception in the nociceptors and spinal

cord. J Pharmacol Exp Ther 291:308–313.

Itoh M, Takasaki I, Andoh T, Nojima H, Tominaga M, Kuraishi Y.

2001. Induction by carrageenan inflammation of prepronociceptin

mRNA in VR1-immunoreactive neurons in rat dorsal root ganglia.

Neurosci Res 40:227–233.

Kanaan SA, Saade NE, Haddad JJ, Abdelnoor AM, Atweh SF, Jabbur SJ,

Safieh-Garabedian B. 1996. Endotoxin-induced local inflammation and

hyperalgesia in rats and mice: a new model for inflammatory pain. Pain

66:373–379.

Kim WG, Mohney RP, Wilson B, Jeohn GH, Liu B, Hong JS. 2000.

Regional difference in susceptibility to lipopolysaccharide-induced neu-

rotoxicity in the rat brain: role of microglia. J Neurosci 20:6309–6316.

Kolanus W, Nagel W, Schiller B, Zeitlmann L, Godar S, Stockinger H,

Seed B. 1996. Alpha L beta 2 integrin/LFA-1 binding to ICAM-1

induced by cytohesin-1, a cytoplasmic regulatory molecule. Cell 86:

233–242.

Kolesnikov YA, Pasternak GW. 1999. Peripheral orphanin FQ/nocicep-

tin analgesia in the mouse. Life Sci 64:2021–2028.

Kurt-Jones EA, Sandor F, Ortiz Y, Bowen GN, Counter SL, Wang TC,

Finberg RW. 2004. Use of murine embryonic fibroblasts to define

Toll-like receptor activation and specificity. J Endotoxin Res 10:419–

424.

Latz E, Visintin A, Lien E, Fitzgerald KA, Monks BG, Kurt-Jones EA,

Golenbock DT, Espevik T. 2002. Lipopolysaccharide rapidly traffics to

and from the Golgi apparatus with the toll-like receptor 4–MD-2–

CD14 complex in a process that is distinct from the initiation of signal

transduction. J Biol Chem 277:47834–47843.

Lindsay RM. 1988. Nerve growth factors (NGF, BDNF) enhance axonal

regeneration but are not required for survival of adult sensory neurons.

J Neurosci 8:2394–2405.

Mansell A, Reinicke A, Worrall DM, O’Neill LA. 2001. The serine pro-

tease inhibitor antithrombin III inhibits LPS-mediated NF-kappaB acti-

vation by TLR4. FEBS Lett 508:313–317.

Medzhitov R, Preston-Hurlburt P, Janeway CA Jr. 1997. A human

homologue of the Drosophila Toll protein signals activation of adaptive

immunity. Nature 388:394–397.



Mika J, Li Y, Weihe E, Schafer MK. 2003. Relationship of pronocicep-

tin/orphanin FQ and the nociceptin receptor ORL1 with substance P

and calcitonin gene-related peptide expression in dorsal root ganglion

of the rat. Neurosci Lett 348:190–194.

Miura Y, Shimazu R, Miyake K, Akashi S, Ogata H, Yamashita Y,

Narisawa Y, Kimoto M. 1998. RP105 is associated with MD-1 and

transmits an activation signal in human B cells. Blood 92:2815–2822.

Miyake K, Shimazu R, Kondo J, Niki T, Akashi S, Ogata H, Yamashita

Y, Miura Y, Kimoto M. 1998. Mouse MD-1, a molecule that is physi-

cally associated with RP105 and positively regulates its expression.

J Immunol 161:1348–1353.

Miyake K, Ogata H, Nagai Y, Akashi S, Kimoto M. 2000. Innate recog-

nition of lipopolysaccharide by Toll-like receptor 4/MD-2 and

RP105/MD-1. J Endotoxin Res 6:389–391.

Mogil JS, Pasternak GW. 2001. The molecular and behavioral pharma-

cology of the orphanin FQ/nociceptin peptide and receptor family.

Pharmacol Rev 53:381–415.

Muroi M, Ohnishi T, Tanamoto K. 2002. Regions of the mouse CD14

molecule required for toll-like receptor 2- and 4-mediated activation of

NF-kappa B. J Biol Chem 277:42372–42379.

Nagai Y, Akashi S, Nagafuku M, Ogata M, Iwakura Y, Akira S,

Kitamura T, Kosugi A, Kimoto M, Miyake K. 2002a. Essential role of

MD-2 in LPS responsiveness and TLR4 distribution. Nat Immunol

3:667–672.

Nagai Y, Shimazu R, Ogata H, Akashi S, Sudo K, Yamasaki H, Hayashi

S, Iwakura Y, Kimoto M, Miyake K. 2002b. Requirement for MD-1

in cell surface expression of RP105/CD180 and B-cell responsiveness

to lipopolysaccharide. Blood 99:1699–1705.

Nguyen MD, Julien JP, Rivest S. 2002. Innate immunity: the missing

link in neuroprotection and neurodegeneration? Nat Rev Neurosci

3:216–227.

Okuda-Ashitaka E, Ito S. 2000. Nocistatin: a novel neuropeptide

encoded by the gene for the nociceptin/orphanin FQ precursor. Pepti-

des 21:1101–1109.

Pettersson LM, Sundler F, Danielsen N. 2002. Expression of orphanin

FQ/nociceptin and its receptor in rat peripheral ganglia and spinal cord.

Brain Res 945:266–275.

Re F, Strominger JL. 2002. Monomeric recombinant MD-2 binds toll-

like receptor 4 tightly and confers lipopolysaccharide responsiveness.

J Biol Chem 277:23427–23432.

Reeve AJ, Patel S, Fox A, Walker K, Urban L. 2000. Intrathecally

administered endotoxin or cytokines produce allodynia, hyperalgesia

and changes in spinal cord neuronal responses to nociceptive stimuli in

the rat. Eur J Pain 4:247–257.

Schromm AB, Lien E, Henneke P, Chow JC, Yoshimura A, Heine H,

Latz E, Monks BG, Schwartz DA, Miyake K, Golenbock DT. 2001.

Molecular genetic analysis of an endotoxin nonresponder mutant cell

line: a point mutation in a conserved region of MD-2 abolishes endo-

toxin-induced signaling. J Exp Med 194:79–88.

Shimazu R, Akashi S, Ogata H, Nagai Y, Fukudome K, Miyake K,

Kimoto M. 1999. MD-2, a molecule that confers lipopolysaccharide

responsiveness on Toll-like receptor 4. J Exp Med 189:1777–1782.

Viriyakosol S, Tobias PS, Kitchens RL, Kirkland TN. 2001. MD-2 binds

to bacterial lipopolysaccharide. J Biol Chem 276:38044–38051.

Visintin A, Mazzoni A, Spitzer JA, Segal DM. 2001. Secreted MD-2 is a

large polymeric protein that efficiently confers lipopolysaccharide sensi-

tivity to Toll-like receptor 4. Proc Natl Acad Sci U S A 98:12156–

12161.

Watkins LR, Wiertelak EP, Goehler LE, Smith KP, Martin D, Maier SF.

1994. Characterization of cytokine-induced hyperalgesia. Brain Res

654:15–26.

Wright SD, Ramos RA, Tobias PS, Ulevitch RJ, Mathison JC. 1990.

CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS

binding protein. Science 249:1431–1433.

Yazawa N, Fujimoto M, Sato S, Miyake K, Asano N, Nagai Y,

Takeuchi O, Takeda K, Okochi H, Akira S, Tedder TF, Tamaki K.

2003. CD19 regulates innate immunity by the toll-like receptor RP105

signaling in B lymphocytes. Blood 102:1374–1380.

Yu B, Wright SD. 1996. Catalytic properties of lipopolysaccharide (LPS)

binding protein. Transfer of LPS to soluble CD14. J Biol Chem

271:4100–4105.

Zarember KA, Godowski PJ. 2002. Tissue expression of human Toll-like

receptors and differential regulation of Toll-like receptor mRNAs in

leukocytes in response to microbes, their products, and cytokines.

J Immunol 168:554–561.



bacterial lipopolysaccharide regulates nociceptin expression in sensory neurons

Documents